Gromov's Mass: A Multifaceted Geometric Measure

• Gromov’s Mass is a multifaceted concept referring to various quantitative measures in geometry, such as filling area, total measure, and isoperimetric mass, each defined within specific contexts.
• Methodologies include combinatorial approaches for discrete fillings, metric measure frameworks to assess total mass, and energy conservation techniques in pseudoholomorphic curve theory.
• Its applications span scalar-curvature comparisons, isometric filling problems, and topological overlap density analysis, offering robust tools for weak-convergence and geometric analysis.

Gromov's Mass is not a standard single invariant in the current literature. In sources associated with Gromov's geometry, the phrase functions, at most, as a contextual label for several distinct quantitative notions of size: filling area under non-shortcut boundary constraints, asymptotic mass of nonnegative-scalar-curvature $3$-manifolds, total measure in metric-measure geometry, symplectic area and energy of pseudoholomorphic curves, and overlap density in topological selection theorems. The common theme is extremal measurement under weak regularity, comparison geometry, or filling constraints rather than a universally accepted scalar invariant (Shioya, 2014, Jauregui et al., 2016, Briggs et al., 19 Feb 2026, Sukhov et al., 2012, Matoušek et al., 2011).

1. Terminological status

Across the cited sources, no notion explicitly named “Gromov’s mass” is defined. In Shioya’s treatment of metric measure geometry, the word mass appears mainly in the ordinary measure-theoretic sense of total mass of a measure and in mass transport, not as a standalone Gromov invariant. In the exposition of Gromov’s pseudoholomorphic-curve method, there is likewise no standalone invariant under that name, and no discussion of current mass in the Federer–Fleming sense. In the discrete filling-area paper, the authors state that they do not explicitly formulate or study a notion called mass; the principal minimized quantity is surface area, and the central invariant is filling area. By contrast, the lower-semicontinuity paper describes its problem as lying “in the spirit of ‘Gromov’s Mass’,” but the concrete object there is Huisken’s isoperimetric mass in dimension $3$ (Shioya, 2014, Sukhov et al., 2012, Briggs et al., 19 Feb 2026, Jauregui et al., 2016).

This suggests that “Gromov’s Mass” is best interpreted contextually. In geometric analysis it may refer to a robust mass notion compatible with weak convergence and scalar-curvature bounds. In filling problems it may refer to an extremal area of a filling. In metric-measure geometry it may refer literally to total measure or, more plausibly, to invariants describing how measure is distributed. In combinatorial-topological settings it may refer to overlap density rather than metric mass.

2. Filling area and isometric fillings

One of the closest Gromovian “mass-type” quantities is filling area. A compact metric surface $M$ is said to isometrically fill a closed metric curve $C$ if $\partial M=C$ and

$$d_M(x,y)=d_C(x,y)\qquad\text{for all }x,y\in \partial M=C.$$

Equivalently, the filling introduces no “shortcuts” between boundary points. In the normalization used for the standard Riemannian circle, the boundary has total length $2\pi$, the round hemisphere has surface area

$2\pi,$

and Gromov’s filling area conjecture asserts that if $M$ is a compact, orientable Riemannian surface that isometrically fills the Riemannian circle of length $2\pi$, then

$3$0

The discrete filling-area paper does not prove the sharp hemisphere bound. It proves instead a universal lower bound valid for arbitrary compact Riemannian $3$1-Lipschitz fillings of a circle of circumference $3$2: $3$3 For isometric fillings, $3$4. For $3$5, this gives

$3$6

The hemisphere, with area $3$7, remains conjecturally optimal, but this bound is a universal obstruction for arbitrary topology and even without orientability assumptions in the lower-bound theorem.

The mechanism is combinatorial. The boundary circle is replaced by the cycle graph $3$8, and fillings are abstract triangulations $3$9 with $M$0. For $M$1, a $M$2-Lipschitz filling satisfies

$M$3

The principal discrete theorem states that if $M$4 is a $M$5-Lipschitz filling of $M$6, then

$M$7

Using balanced triangulations of PL metric surfaces and approximation by piecewise-flat surfaces, this quadratic combinatorial lower bound is transferred to the continuous area estimate. In this setting, “mass” is therefore best understood as a filling-area quantity: a two-dimensional filling size constrained by preservation of the boundary metric (Briggs et al., 19 Feb 2026).

3. Isoperimetric mass and lower semicontinuity under $M$8 convergence

In asymptotically flat scalar-curvature geometry, the mass notion closest to a robust “Gromov-style” invariant is Huisken’s isoperimetric mass. For a sequence of asymptotically flat $M$9-manifolds of nonnegative scalar curvature with outermost minimal boundary, converging in the pointed $C$0 Cheeger–Gromov sense to an asymptotically flat limit space, the lower-semicontinuity theorem states that the total mass of the limit is bounded above by the liminf of the total masses of the sequence. The result is specific to dimension $C$1, and its significance is precisely that ADM mass is hard to control directly under $C$2 convergence, whereas isoperimetric mass remains meaningful because it depends only on areas and volumes (Jauregui et al., 2016).

The same work proves that for smooth asymptotically flat $C$3-manifolds with nonnegative scalar curvature, isoperimetric mass agrees with ADM mass. This identification makes isoperimetric mass a low-regularity substitute for ADM mass: it is defined from volume and area alone, agrees with the classical mass in the smooth setting, and is stable enough to survive weak convergence. The quasilocal version measures the excess of actual enclosed volume over Euclidean isoperimetric comparison volume, normalized by boundary area.

The geometric role of this notion is structural rather than terminological. It provides a concrete answer to a “Gromov-style” problem: how to encode mass in a way that is geometric, compatible with scalar curvature, and robust under weak convergence. The proof uses Huisken’s relative volume monotonicity, a modified weak mean curvature flow, and a freezing procedure for components whose perimeter drops below $C$4. The resulting theory shows that lower semicontinuity of mass can be obtained from coarse $C$5 data without direct access to curvature or first-derivative asymptotics (Jauregui et al., 2016).

4. ADM mass as cubical defect

A different mass interpretation appears in the cubical formula for ADM mass on asymptotically flat $C$6-manifolds. Replacing large coordinate spheres by large coordinate cubes, the mass is expressed through geometry on faces and edges: $C$7 Equivalently, with $C$8 the dihedral angle,

$C$9

Here the face term records mean curvature and the edge term records deviation from right-angle Euclidean cube geometry. The same mass can also be written as a sliced Gauss–Bonnet defect: $\partial M=C$0 where $\partial M=C$1 is the coordinate-square slicing curve in the plane $\partial M=C$2, $\partial M=C$3 is its geodesic curvature in that plane, and $\partial M=C$4 is the sum of its four turning angles.

This realizes mass as an integrated angle defect. In Euclidean space, each face is flat, each dihedral angle is $\partial M=C$5, and the defect vanishes. The formula is explicitly related to Gromov’s scalar-curvature comparison theory for cubic Riemannian polyhedra: mass measures the asymptotic failure of a large coordinate cube to be Euclidean, encoded by face mean curvature and edge-angle deviation. Under nonnegative scalar curvature, the positive mass theorem yields an asymptotic integral cubical inequality

$\partial M=C$6

up to the stated error term. In this sense, ADM mass becomes a scalar measure of total cubical defect (Miao, 2019).

5. Euclidean endpoint rigidity below the mass scale

The Euclidean endpoint $\partial M=C$7 rigidity theorem sharpens the relation between mass and asymptotic decay. If $\partial M=C$8 is a smooth complete metric on $\partial M=C$9 with non-negative scalar curvature and

$$d_M(x,y)=d_C(x,y)\qquad\text{for all }x,y\in \partial M=C.$$0

then $$d_M(x,y)=d_C(x,y)\qquad\text{for all }x,y\in \partial M=C.$$1 is isometric to Euclidean space. No derivative decay is assumed. The point is that $$d_M(x,y)=d_C(x,y)\qquad\text{for all }x,y\in \partial M=C.$$2 is precisely the scale below the first nontrivial asymptotic mass term; at the exact $$d_M(x,y)=d_C(x,y)\qquad\text{for all }x,y\in \partial M=C.$$3 scale there exist smooth non-flat metrics with $$d_M(x,y)=d_C(x,y)\qquad\text{for all }x,y\in \partial M=C.$$4 such that

$$d_M(x,y)=d_C(x,y)\qquad\text{for all }x,y\in \partial M=C.$$5

near infinity (You et al., 13 Jun 2026).

This theorem is a rigidity statement rather than a direct mass formula. It shows that the zero-mass rigidity phenomenon survives even when the usual ADM integral is unavailable because only $$d_M(x,y)=d_C(x,y)\qquad\text{for all }x,y\in \partial M=C.$$6 asymptotics are assumed. The proof uses a positive Green function for a divergence-form operator associated with $$d_M(x,y)=d_C(x,y)\qquad\text{for all }x,y\in \partial M=C.$$7, a moving-dipole asymptotic expansion at infinity, and the Agostiniani–Mazzieri–Oronzio monotone functional $$d_M(x,y)=d_C(x,y)\qquad\text{for all }x,y\in \partial M=C.$$8. The critical structural fact is that the first variation of the averaged $$d_M(x,y)=d_C(x,y)\qquad\text{for all }x,y\in \partial M=C.$$9-functional annihilates every Euclidean dipole mode, allowing endpoint control despite the lack of a fixed ADM-type asymptotic coefficient.

Within the broader landscape of “Gromov’s mass,” this result isolates the exact Euclidean threshold beyond which no nontrivial positive-mass geometry can persist. A plausible implication is that, in dimension $2\pi$0, rigidity is more robust than the classical differentiable definition of mass itself (You et al., 13 Jun 2026).

6. Total mass and concentration in metric-measure geometry

In metric-measure geometry, the nearest literal meaning of mass is the total measure $2\pi$1. Shioya defines an mm-space as a triple

$2\pi$2

where $2\pi$3 is a complete separable metric space and $2\pi$4 is a Borel probability measure on $2\pi$5. Thus, in the standard normalization used throughout the theory,

$2\pi$6

The introduction explicitly remarks that in Gromov’s original treatment the measures of mm-spaces are not necessarily probability, but the proofs extend easily to non-probability mm-spaces. In that broader setting, the most literal “mass” of a space is therefore its total finite measure (Shioya, 2014).

Yet the actual Gromovian content of the theory lies less in the scalar $2\pi$7 than in invariants describing how the unit mass is distributed. The principal objects are observable diameter, concentration function, separation distance, box distance, observable distance $2\pi$8, measurements $2\pi$9, and pyramids $2\pi,$0. These are not called mass, but they are the main quantitative descriptors of measure concentration and observable geometry.

For example, observable diameter studies the largest partial diameter of $2\pi,$1 over all $2\pi,$2-Lipschitz maps $2\pi,$3, while $2\pi,$4-measurements

$2\pi,$5

record all push-forward probability measures obtained by $2\pi,$6-Lipschitz observations into $2\pi,$7. Distance matrix distributions determine the mm-space up to mm-isomorphism, and a sequence is a Lévy family precisely when it concentrates to a one-point mm-space in $2\pi,$8. In this setting, “mass” is best read either literally as total measure or structurally as the full observable distribution of measure (Shioya, 2014).

7. Symplectic area, energy, and overlap density

In Gromov’s theory of pseudoholomorphic curves, the operative quantity is symplectic area rather than any separately named mass. For a $2\pi,$9-complex curve $M$0, the paper defines

$M$1

For discs, it also defines the energy

$M$2

and states the identity

$M$3

Under bubbling, the relevant conservation law is

$M$4

so total energy or area is redistributed into the limiting map and the bubbles. This is the clearest “mass-conservation” statement in the compactness theory, but it is an area/energy principle rather than a separate formal mass invariant (Sukhov et al., 2012).

A different non-metric notion appears in Gromov’s method for heavily covered points. There the relevant quantity is overlap density: the fraction of all $M$5-simplices spanned by a point set, or by a continuous image of the simplex, that pass through a common point. The central constant $M$6 is the optimal guaranteed asymptotic overlap fraction. Gromov’s method yields

$M$7

and in dimension $M$8 the refined bound

$M$9

hence

$2\pi$0

Here the “mass” is neither physical nor metric; it is the normalized coverage count of simplices through a point. This is a topological overlap mass, forced by coboundary expansion and the noncontractibility of a cycle in the space of cocycles (Matoušek et al., 2011).

In this broader synthesis, “Gromov’s Mass” names no single object. It designates a family of Gromovian size notions—filling area, asymptotic mass, total measure, symplectic area-energy, and overlap density—whose common purpose is to measure geometric or topological obstruction under comparison, filling, or weak-convergence constraints.